Senior Generative & Agentic AI Risk Specialist

UOB; Singapore; On-site; SGD 180,000 - 280,000; Full time

Job summary

United Overseas Bank Limited (UOB) is seeking an experienced AI risk governance professional focused on Generative and Agentic AI risk assessment and oversight. The role executes end‑to‑end use case risk assessments, monitors frontier AI technologies, conducts applied research, and communicates decision‑oriented insights to senior management.

Prototypes and PoCs may be developed to support validation and testing.

Qualifications

• Experience: 5–10 years of relevant experience in AI or AI and data governance, preferably in banking or financial services.
• Independent validation and business‑as‑usual risk assessment: Design and execute independent validation activities including evaluation design, benchmark construction, test‑suite development, scenario testing, uncertainty analysis and monitoring effectiveness reviews.
• Third‑party risk assessment: Knowledge of risk assessment processes applicable to externally provided AI vendors, models, platforms, services, or data.
• AI risk controls and guardrails: Knowledge of relevant AI control and guardrail patterns, operating‑effectiveness testing, and supporting evidence for Generative AI and Agentic AI.
• Experience conducting adversarial testing, red teaming, evaluation benchmarking, model behavior analysis and AI control effectiveness testing across Generative AI and Agentic AI.
• Market research and senior‑management communication: Ability to conduct structured horizon scanning and communicate concise, decision‑oriented findings and recommendations to senior management.
• Hands‑on experience developing prototypes or proof‑of‑concepts using modern AI frameworks such as LangChain, LangGraph, Semantic Kernel, AutoGen, CrewAI, OpenAI, Anthropic, Azure AI, AWS Bedrock or equivalent platforms. Ability to understand and assess architectures involving RAG, agents, tool use, MCP, orchestration frameworks, vector databases, model gateways and AI observability tooling.
• Applied research: Ability to critically assess research, standards, and industry practices; design or review experiments; and translate findings into validation methods, evaluation metrics, test suites, and internal guidance for Generative AI and Agentic AI.
• Communication and effective challenge: Ability to communicate technical and risk findings clearly to technical and non‑technical stakeholders.
• Delivery model: Self‑directed individual contributor capable of working across multidisciplinary teams.

Responsibilities

• Perform independent review and effective challenge of AI use cases, covering Generative AI and Agentic AI. Execute business‑as‑usual risk assessment activities, including inherent risk assessment, control effectiveness assessment, residual risk assessment, remediation tracking, and governance reporting. Apply third‑party risk assessment processes where external providers or solutions are involved.
• Scan the market and research the frontier for emerging AI technologies, attack techniques, incidents, regulatory developments, standards, and industry practices. Assess the implications for the bank and present clear, decision‑oriented findings to senior management.
• Develop proof of concepts, prototypes, evaluation environments and reference implementations for Generative AI and Agentic AI solutions to support independent validation, risk assessment, control testing, and effective challenge. Understand key architectural components including models, orchestration frameworks, agents, tools, retrieval systems, APIs, data pipelines and deployment patterns.
• Conduct applied research specific to Generative AI and Agentic AI to strengthen validation methodologies, evaluation approaches, metrics, test suites, and internal guidance.
• Apply domain knowledge of AI risk evaluations, covering relevant threats and failure modes such as prompt injection, jailbreak, data poisoning, distribution shift, unsafe tool use, and agentic failure modes.
• Apply working knowledge of relevant AI risk frameworks and industry guidance, including GenAI/Agentic AI papers from MAS, IMDA, ABS, OWASP, SAFR and so on.
• Implement Hands‑on experience developing prototypes or proof‑of‑concepts using LangChain etc. Understand architectures involving RAG, agents, orchestration, vector databases, and AI observability tooling.
• Applied research: ability to assess research, standards, and industry practices; design experiments; translate findings into validation methods and internal guidance.
• Communication and effective challenge: to convey technical and risk findings to diverse stakeholders.
• Delivery: self‑directed IC working across multidisciplinary teams.

Skills

Generative AI

Agentic AI

AI risk governance

Independent validation

Threat modelling

Horizon scanning

Education

Bachelor’s in CS / Data Science / Math

Postgraduate in ML / AI safety (advantage)

Tools

LangChain

LangGraph

Semantic Kernel

OpenAI

Job description

United Overseas Bank Limited (UOB) is seeking an experienced AI risk governance professional focused on Generative and Agentic AI risk assessment and oversight. The role executes end‑to‑end use case risk assessments, monitors frontier AI technologies, conducts applied research, and communicates decision‑oriented insights to senior management.

Prototypes and PoCs may be developed to support validation and testing.